282s to generate12,730 output tokens
← All AI models
OpenAI · GPT-5
GPT-5
Previous flagshipText / reasoning model
OpenAI’s mainline reasoning model and the successor to GPT-4. Still real, still strong — now one tier below GPT-5.5.
Yes — GPT-5 is a real OpenAI model.
Released 2025 by OpenAI, GPT-5 is part of the GPT-5 family. It's available through OpenAI's API, and we run it on PromptFrenzy — the examples below are real, unedited generations it produced.
Maker
OpenAIReleased
2025
Input price
$1.25 / 1M
Output price
$10 / 1M
What is GPT-5?
GPT-5 is the model most people mean when they say “GPT-5”: OpenAI’s general-purpose reasoning flagship that replaced the GPT-4 line. It runs with adjustable reasoning effort and remains a heavyweight on open-ended, multi-step tasks. As of mid-2026 it sits just behind GPT-5.5 as OpenAI’s mainline model, and it’s one of the five models in our Dream Home showdown — where it produced a complete, original voxel build from a single prompt.
Strengths
- ✓Strong on long, open-ended reasoning tasks — held its own against newer flagships in our showdowns.
- ✓Adjustable reasoning effort lets you trade latency for depth.
- ✓Cheaper per token than the newer GPT-5.5 while staying close in quality.
Trade-offs
- ✗Superseded as OpenAI’s mainline model by GPT-5.5.
- ✗High-effort reasoning runs are slow — we saw ~4–5 minutes on heavy generative tasks.
- ✗Reasoning tokens bill as output tokens, so a “short” answer can cost more than it looks.
What GPT-5 actually produces
Real, unedited outputs from GPT-5 in our model showdowns — same brief, same constraints as every other model, one shot each.
“I want to work where stone, water, and sky meet: a quiet, light-filled studio cantilevered over a cold cove with a glass roof and a little tree to keep me company. It’s a place for writing, tinkering, and slipping down a ladder for a swim at sunset.”
290s to generate23,056 output tokens
See GPT-5 compared head-to-head
Compare every model side-by-side
See how GPT-5 stacks up against the other frontier models on the same brief.
Browse all model showdowns →